Abstract:
Lung cancer subtyping, particularly differentiating adenocarcinoma (ADC) from squamous cell carcinoma (SCC), is paramount for clinicians to develop effective treatment strategies. In this study, we aimed: (i) to discover volatile organic compound (VOC) biomarkers for precise diagnosis of ADC and SCC, (ii) to investigated the impact of risk factors on ADC and SCC prediction, and (iii) to explore the metabolic pathways of VOC biomarkers. Exhaled breath samples from patients with ADC (n= 149) and SCC (n= 94) were analyzed by gas chromatography-mass spectrometry. Both multivariate and univariate statistical analysis method were employed to identify VOC biomarkers. Support vector machine (SVM) prediction models were developed and validated based on these VOC biomarkers. The impact of risk factors on ADC and SCC prediction was investigated. A panel of 13 VOCs was found to differ significantly between ADC and SCC. Utilizing the SVM algorithm, the VOC biomarkers achieved a specificity of 90.48%, a sensitivity of 83.50%, and an area under the curve (AUC) value of 0.958 on the training set. On the validation set, these VOC biomarkers attained a predictive power of 85.71% for sensitivity and 73.08% for specificity, along with an AUC value of 0.875. Clinical risk factors exhibit certain predictive power on ADC and SCC prediction. Integrating these risk factors into the prediction model based on VOC biomarkers can enhance its predictive accuracy. This work indicates that exhaled breath holds the potential to precisely detect ADCs and SCCs. Considering clinical risk factors is essential when differentiating between these two subtypes.
摘要:
肺癌分型,特别是区分腺癌(ADC)和鳞状细胞癌(SCC),对于临床医生制定有效的治疗策略至关重要。在这项研究中,我们的目标是:(I)发现VOC生物标志物,用于ADC和SCC的精确诊断,(ii)调查风险因素对ADC和SCC预测的影响,和(iii)探索VOC生物标志物的代谢途径。通过气相色谱-质谱(GC-MS)分析了ADC(n=149)和SCC(n=94)患者的呼气样本。多变量和单变量统计分析方法均用于鉴定VOC生物标志物。基于这些VOC生物标志物开发并验证了支持向量机(SVM)预测模型。研究了危险因素对ADC和SCC预测的影响。发现一组13个VOC在ADC和SCC之间存在显着差异。利用SVM算法,VOC生物标志物的特异性达到90.48%,灵敏度为83.50%,训练集上的AUC值为0.958。在验证集上,这些VOC生物标志物的敏感性和特异性分别为85.71%和73.08%,AUC值为0.875。临床危险因素对ADC和SCC预测具有一定的预测能力。将这些风险因素整合到基于VOC生物标志物的预测模型中可以提高其预测准确性。这项工作表明,呼出气具有精确检测ADC和SCC的潜力。在区分这两种亚型时,考虑临床风险因素至关重要。